I'd like to discuss the subject of Variable fonts. Are we sure they are the future, or is it just something we are trying to push on the market, without the customers really wanting/needing it? I've spoken to a number of graphic designers who say they don't want to drag sliders when setting type, they want pre-made weights. Myself, I see each weight as a "work of applied art" (perhaps a bit pretentious), carefully selected and refined, rather than a font being some kind of interactive software.
Most designers I know have heard about variable fonts but don't know that much about them. Only recently have they begun to run into them.
When designers buy fonts, they typically purchase one or two weights for a particular project. Purchasing a more expensive variable font that contains more than the weights they need is often outside the project budget.
In addition, they're concerned (and not without reason) that variable fonts won't print correctly or their software won't support the fonts.
Even so, I think there's a future for variable fonts for the simple reason that most type designers rely on interpolation to create intermediate weights anyway. As long as they're doing that, it's sort of a no-brainer to save out a variable font too, even if they're not widely used.
In addition, as old Type 1 font families soon become unusable, there might be a push to buy newer replacements. Rather than buy a dozen fonts to get the whole family, people just might be tempted to buy the variable versions to save a bit of money.
Eventually, there will be a critical mass of them available, graphics software will do a better job supporting them, and they'll gain acceptance. Of course, I could easily be wrong.
Mostly agree, but I don't think any foundry offers a variable version separate from (and cheaper than) the whole family. How I offer it, and most foundries that I know of too, is that anybody who buys the whole family also gets a copy of the variable font along with all the static files.
Of the latter, how many might be willing to pay a bit more (vs $ x N styles) to get a variable font? Is there a model here that might lead to net increase in revenue?
(Not a rhetorical question: I don’t make or sell fonts, so don’t have any first hand knowledge on this.)
I suspect that, as @Peter Constable mentioned, there could be a viable market in selling variable fonts at a bit of a discount to people who would usually only pick and choose the fonts they needed from a larger family. I'm only speculating, though.
I think it's a mistake to assume that all purchasers of variable fonts are wanting "the entire family". They may only want a few custom styles, for example. In fact, most customers of Proxima Nova only purchase a few styles. Forcing people to pay the same as a full 48-style family would exclude most of the existing market for the static version, severely limiting the market for the variable version.
- what to charge if one sells ONLY the variable version
- what to charge if one sells BOTH variable and non-variable versions of the same family
In the former case, one likely wants to price in a way that makes the variable font attractive to those who don’t want to pay the full-family price. Also, probably to maximize revenue (or profit, which is a bit more tricky, but the marginal cost of additional units sold is low, aside from support costs.
There can be endless tweaks to both the static and variable versions for high-quality work — some typeface designs more than others. But my point was the relative ease of producing a variable font from interpolation is something of a no-brainer for most, which I suggested might help ensure the continuation and eventual success of variable fonts.
I don't think we disagree, however, that the highest-quality work isn't quite that simple. Perhaps you could provide some examples of how that's the case. I'm always eager to learn. As I mentioned, I'm primarily an art director who uses type. Designing type and building fonts is an increasingly serious after-hours thing for me, so more information is always welcome.
My thinking is to price the variable version low enough that it will be attractive to anyone who wants more than a few styles. (If they only need a few and want to save money, the static version still exists.) Only a tiny percent of customers purchase an entire static family. I don't want to limit the market to that subset. The bulk of my customers purchase 1-6 styles from the 48 available. I want the price to be attractive to those customers. My hunch is that this will allow the variable version to bring in at least as much income as the static version. I don't know yet if I'm correct. There are still technical issues with variable fonts that may be a factor for users in the short term more than price. Time will tell and it's still early. But I honestly don't think variable fonts will catch on if they are only available at a full static family price.
Just as OpenType fonts reduced the “price per glyph”, VF will reduce the “price per weight”. But perhaps the best metric for us is “price per master”.
Purchases are presumably also migrating to the VF from individual styles.
For a general overview of variable fonts and how they work, my 2016 article may still be useful.
One reason for this is that significant differences exist between the italic and roman designs (number and position of anchor points and unique glyph designs) that make it impractical to interpolate them without many workarounds, design compromises, and intermediate instances. A simple slanted axis would be easy enough, but not a more traditional custom-designed italic.
Even with separate variable roman and italic fonts in the same typeface, both will appear together in a Mac's font list (you specifically asked about Macs). But the roman and italic will have separate sliders.
Thank you for the details. I understand, instances (and masters) should be linked properly italic to roman as a family, and the axes should be equal, even on different variable fonts, of course.
For example, if the both files have three axes (Weight, X-height, Contrast), and ... I need to add an additional (fake) Italic axis (with 0 for roman and 1 for italic) to the both files, even if the lowercase have a different construction? Does it needed for compatibility?